Table 5 Statistical comparison of the gravity anomaly differences between the marine gravity models and the shipborne gravity anomalies at the checkpoints at different depths (unit: mGal).

From: Refining satellite Altimetry-Derived gravity anomaly model with shipborne gravity using multilayer perceptron neural networks

Depths/km

Number of checkpoints

Modle

MAX

MIN

MEAN

STD

RMS

MAE

(0, 1]

17,861

MLP_GRA

26.33

-29.10

-0.62

2.55

2.62

1.87

SDUST2022GRA

22.37

-34.43

0.42

2.84

2.87

2.09

SIO_V32.1

30.85

-30.17

-0.13

3.04

3.04

2.17

DTU21GRA

28.02

-33.54

0.36

2.78

2.80

2.01

(1, 2]

9251

MLP_GRA

34.42

-44.33

0.13

2.96

2.96

2.00

SDUST2022GRA

31.04

-41.98

-0.05

3.71

3.71

2.58

SIO_V32.1

32.29

-43.40

0.14

3.80

3.80

2.59

DTU21GRA

32.31

-40.59

-0.08

3.83

3.83

2.62

(2, 3]

7191

MLP_GRA

32.35

-25.11

0.25

3.22

3.23

2.13

SDUST2022GRA

23.90

-33.59

-0.44

3.54

3.57

2.49

SIO_V32.1

29.21

-34.69

-0.16

3.75

3.76

2.58

DTU21GRA

23.68

-36.54

-0.37

3.53

3.55

2.46

(3, 4]

10,756

MLP_GRA

31.78

-27.11

0.18

2.75

2.75

1.92

SDUST2022GRA

37.33

-31.12

-0.35

3.07

3.09

2.19

SIO_V32.1

37.03

-28.17

-0.06

3.21

3.21

2.25

DTU21GRA

36.24

-27.01

-0.26

3.05

3.06

2.13

(4, +∞)

2228

MLP_GRA

27.52

-30.37

1.10

3.81

3.96

2.70

SDUST2022GRA

25.04

-42.36

-0.42

4.42

4.44

2.86

SIO_V32.1

27.14

-41.50

0.54

4.39

4.42

2.88

DTU21GRA

24.63

-41.15

0.13

4.36

4.36

2.77